一种基于分数阶微分电压-容量曲线的锂离子电池健康状态估计新方法

Abstract Accurate estimation of state of health (SOH) of lithium-ion batteries is crucial to ensure that the battery management system stably runs. Extraction of characteristic parameters (CPs) is key for accurate prediction of SOH. Traditional methods for extracting CPs have certain limitations like a small number of CPs and having some difficult for extracting features. To address the above issues, this study combines the Caputo fractional derivatives theory with the voltage-capacity curve, introducing the fractional-order differential voltage-capacity curve for CPs extraction. Additionally, this article introduces v -support vector machine, elastic net, and proposes closed-loop gaussian process regression, utilizing a fusion model algorithm to combine these three models into one fusion model, enhancing the SOH estimation accuracy. Finally, we did different sets of comparison experiments: using the CPs extracted in this paper as inputs to different models to verify that the estimation accuracy of the fusion model algorithm is higher than that of other traditional methods; using different CPs as the inputs to the same model to verify the validity of the CPs extracted in this paper. The experimental results show that both the fusion model and the CPs proposed in this paper have better performance in estimating SOH.